ICD deactivation in patients approaching the end of life. Are we getting it right?

Background

In patients approaching end of life with implantable cardiac defibrillators (ICDs) it is critical to deactivate ICD therapies to avoid futile shocks.

Our NHS trust has developed pathways for ICD deactivation to improve quality of care.

Aim

Audit requests for ICD deactivation within our trust from January 2022 to January 2024 and identify areas for service improvement.

Method

All deactivation requests (including post-mortem) are made through an electronic system. We conducted a retrospective case note review.

Results

73 patients had an ICD deactivation request. Median age was 79.

8/73 had deactivation requests after death and therefore an active device at death. Of these, 4/8 died in hospital. 3/4 that died in hospital had DNAR (do not attempt resuscitation) orders and their death was anticipated. These 3 patients clinical notes had wrongly listed the device as a pacemaker. 1/4 died during active management on critical care. The remaining 4/8 unexpectedly died at home.

26/73 patients received any therapy from the device following implantation and 13/73 patients received shock therapy.

10/26 who had received therapy received any therapy in the 3 months before death. 5 received any therapy 24 hours before death.

3 patients received shock therapy at the time of death. A further two received therapy in the final 24 hours of life. 4/5 patients had a DNAR at the time.

In the 65/73 patients with a deactivation request before death, the median (IQR) time from deactivation to death was 5 days (0.25-48.5). Median (IQR) time to deactivation following request was 0(0-0) days and range 0-3 days.

51/65 requests were made for hospital inpatients, 12/65 were initiated in cardiology outpatients, 2/65 from primary care. Of the 51 hospital inpatients 37 were from medicine, 13 cardiology and 1 critical care.

27 patients had their device deactivated alongside advanced care planning. In these 27 patients the median (IQR) time to death was 252 days (116-533).

Discussion

ICDs were deactivated within a short time frame and no shock therapy occurred awaiting deactivation. When this decision was taken out of hours magnet therapy was effective.

In the 3 patients that died with a DNAR but active device there was a failure to identify the type of device. It is important to reinforce educational message about ICDs to clinical teams and investigating electronic solutions to flag ICDs.

Our data showed two peaks of deactivation. An early group from a joint cardiac physiologist and consultant clinic (frequently in patients approaching ICD generator replacement) and a late deactivation group of hospital inpatients.

Most late deactivation patients were receiving care outside of cardiology when their deactivation decision was made. This was often close to death.

Few patients received device therapy in the months preceding deactivation and death suggesting device therapy will not be a clue to a patient approaching the end of life.